[VarHyid]

Although the topic title speaks for itself, we won't get anywhere without some more detailed info. I wasn't sure whether to put this into the OS section, but I'm not 100% sure if it's really an OS problem.

First off, buying a new PC ASAP isn't an option (I hope I can get a better answer than that here ), trying out different hardware in different PC isn't possible either. Let's get to the details.

What the f*** is happening?
Every time I try to change either screen resolution or color depth, the machine crashes and almost instantly reboots. The reboot part isn't the most unusual problem because it's set to auto-reboot on crash so obviously I'd like to know why does it crash in the first place. It happens as well with games as regular attempt to change the res. via OS' screen properties.

How does it crash exactly?
After the resolution changes (it really does) in about 2 seconds everything freezes (sound loops for a short moment) and bang goes the reboot as if I would press restart. No blue screen, no error message just either a black screen and/or some artifacts on the top of the screen, depending on the resolution it may be 1-2 lines up to 10% of the screen - the lower the res. the more artifacts, almost as if the GFX-card would try to show them and manage only to do a particular amount (hence - more at lo-res).
Also the time varies, from 1s to even 3s. It looks as if it would be unable to "hold" the new resolution for a long time.

Are there any exceptions? Does it always happen?
That's the funny thing. Sometimes it doesn't. I can't exactly tell why and under what conditions, but it looks like it needs time to "adapt" to the new resolution before it saves the new settings ("decides" to run with them). Therefore, the only way I've managed to change the resolution/color depth (after installing new drivers, for example) is to almost let the 15s counting down after the change almost to run out and then click OK. This way it stays and after a normal restart it keeps it, once I try to hit OK right after "apply" - crash and we're back at the previous res.
Also, I have only 256MB RAM so most of the time the OS has to use physical memory and I observed (can't be 100% sure about that, though) that when I start a game while the system has been up for a long time and run many other memory consuming applications in the background and while the game changes the resolution, the HDD has to "take" some more of the paging file, it may not crash as if the required couple of seconds when it would pass and then it's able to keep the settings, just as in case of changing res. from the OS.
I know - sounds crazy.

What can NOT be the problem?
I'm 100% sure that it's NOT:
1. The graphic card (physically) - it started to happen years ago (haven't played games in a long time so it didn't really bothered me) with a GeForce 2 GTS that died recently, the new one is a GeForce 3 and exactly the same thing happens the same way.
2. The monitor - beside the fact that I doubt a monitor would be able to force the system to reboot, I've started with a 4:3 CRT screen years ago (also died quite recently) and now have a 16:10 LCD screen - same thing here.
3. Video cable - also, can't imagine how it should work, but same problem via the old D-Sub cable and now the new DVI-D connection.
4. Overheating - it happens while changing to a higher AND to a lower resolution, as well right after first start in the morning and after hour of work and even, as mentioned earlier, it looks like sometimes it even works "on stress".
5. 3D graphics or other advanced visual... things - I can run windowed 3D games and nothing happens, also the DX-Diag test work... except of course the full-screen one that changes the resolution - or... it works for a second or two.
6. Dust - PC is cleaned every couple of months, have changed the thermal paste on both the CPU and GPU (on the old card).
7. Power consumption - if it was a power problem there would be no way of overcoming the problem and the PC should crash when running a windowed game when the GFX card takes more power, but even plugging a second HDD doesn't make any difference.

Beside that, I'm... quite sure it's not:
1. OS - Started with the old Windows 98 SE, happens at Windows 2000 Professional as well.
2. Memory - although I've had some blue screens while running too many memory consuming processes, MEMtest doesn't show any errors at all.

What CAN be the problem? What did I do when it started?
As for what did I do when it started - good question, next question. If I knew this I probably wouldn't have this problem.
What I think it may be:
1. Memory - although MEMtest doesn't show any problems - it's RDRAM, FFS, it's old as [bleep] (~8 years), it's non-ECC, I've bought 2 x 256 and tired to run: only them (+2 terminators (or how are these empty ones called) on the empty slots... that's RDRAM) - unstable/crash/blue screen/doesn't even boot sometimes, these and 2 of my old ones making it 640 MB - worked one day, the other didn't always, next day didn't boot at all, then once did... etc. So maybe that's what's f***ed up.
2. Video drivers - tried a lot of them with the old card always updated when new ones came out up to the... umm... 7x.xx detonator (can't recall now) - always made it better, then just couldn't install better detonators as they didn't support my card anymore. BUT, the GF3 ran as well with the old ones from the GF2, then some new 8x.xx ones and now I've got the best possible before that can be installed which is ForceWare 93.71. Although the problem happened with any driver, maybe it's not the memory consumption that made the 2 games I've checked run before, but the "worse" drivers? On the other hand - even then, I couldn't change the desktop resolution without problems.
3. DirectX - that was my most probable theory. When I bought the PC, AFAIR DX8.1 was the recent one (or was it DX7?), anyway I remember having some issues with one update... years ago, I think it was DX9.0c, the one I'm on right now. I remember even using the DX-eradicator to downgrade to 9.0 (or 9.0a...b...?), but unfortunately, I can't remember exactly what was the problem. Anyway, I've tried to downgrade to 9.0b yesterday (with the old trick of changing the version number in the registry to force it to "update") and didn't help either. Still, maybe someone's sure that's it...?
4. Motherboard - that's where I'm green. I'm not into electronics so I have no idea what could break there or cause this problem and I haven't updated the BIOS so that's up to you, Experts - is it possible?

What's the config?
Can be found in my profile, but in a nutshell, the most relevant parts are:

ECS P4ITA i850
P4 Willamette 1.3 GHz
256 MB RDRAM
Leadtek GeForce 3
DirectX 9.0c
ForceWare 93.71

I know, an ancient "technology". Don't play games at all, wanted to run 2 old ones to kill some time, but got the old resolution changing "reminder" again.

[Advertisements]

Don't knock the technology - RDRAM is good stuff, if you can find it, and afford it - it was expensive then, outrageous now - but my i850 Gigabyte board with 2 x 512Mb 1066 RDRAM is still running strong - knock on forehead/wood.

So you have replaced the monitor and graphics card. You tested the RAM (did you let it run for several passes?) using MemTest. You might try Windows Memory Diagnostic - see the easy to follow instructions under Quick Start Information.

Disconnect any unneeded devices, especially external USB devices.

7. Power consumption - if it was a power problem there would be no way of overcoming the problem and the PC should crash when running a windowed game when the GFX card takes more power, but even plugging a second HDD doesn't make any difference.

Bad assumption. That would be the scenario for a seriously under-rated power supply, and hard drives consume very little (typically <20W when spinning up and <10W when idle) compared to other devices, especially graphics cards - which can easily out-pace CPUs. A stressed or failing supply, on the other hand, teeters on the threshold of stable/unstable operation, often causing very irregular (and difficult to troubleshoot) behavior in the devices it supplies power to, including the motherboard, while "appearing" to work fine.

Therefore, I ALWAYS start troubleshooting hardware by ENSURING I am feeding it quality power - most easily done by swapping in a known good PSU with sufficient horsepower. I see no mention of your PSU or voltages. What are the specs of your PSU? Is this the original? Has it been tested, or have you tried another? You MUST verify your power is (1) ample, and (2) stable (stays within ATX Form Factor standards - 5 or 10% depending on voltage) at ALL loads. It is a common mistake to ignore power requirements when adding or upgrading (or troubleshooting) hardware. This is especially critical when upgrading power hungry graphics cards.

Below is my canned text on testing PSUs, then another text on calculating power requirements and choosing a quality upgrade.

To properly and conclusively test a power supply unit (PSU), it must be tested under various realistic "loads" then analyzed for excessive [url="http://en.wikipedia.org/wiki/Ripple_(electrical)""]ripple[/url] and other anomalies. This is done by a qualified technician using an oscilloscope or power analyzer - sophisticated (and expensive) electronic test equipment requiring special training to operate, and a basic knowledge of electronics theory to understand the results. Therefore, conclusively testing a power supply is done in properly equipped electronic repair facilities.

Fortunately, there are other options. I keep a FrozenCPU Ultimate PSU Tester in my tool bag when I am "in the field" and don't have a good spare power supply to swap in. While not a certain test, they are better than nothing. The advantage of this model is that it has an LCD readout of the voltage. With an actual voltage readout, you have a better chance of detecting a "failing" PSU, or one barely within specified ATX Form Factor Standard tolerances. Lesser models use LEDs to indicate the voltage is just within some "range". These are less informative, considerably cheaper, but still useful for detecting PSUs that have already "failed". Newegg has several testers to choose from. All these testers contain a "dummy load" to fool the PSU into thinking it is connected to a motherboard, and therefore allows the PSU to power on, if able, without being attached to a motherboard - great for testing fans, but again, it is not a true load or suitable for conclusive testing.

As mentioned, swapping in a known good supply is a tried and trued method of troubleshooting used for centuries, even by pros. Remove the "suspect" part and replace with a "known good" part and see if the problem goes away.

I do not recommend using a multimeter to test power supplies. To do it properly, that is, under a realistic load, the voltages on all the pins must be measured while the PSU is attached to the motherboard and the computer powered on. This requires poking (with some considerable force) two hard and sharp, highly conductive meter probes into the main power connector, deep in the heart of the computer. One tiny slip can destroy the motherboard, and everything plugged into it. It is not worth the risk considering most multimeters, like plug-in testers, do not measure, or reveal any unwanted and potentially disruptive AC components to the DC voltages.

And remember, anything that plugs into the wall can kill. Do not open the power supply's case unless you are a qualified electronics technician. There are NO user serviceable parts inside a power supply.

***

Use the eXtreme PSU Calculator Lite to determine your power supply unit (PSU) requirements. Plug in all the hardware you think you might have in 2 or 3 years (extra drives, bigger or 2nd video card, more RAM, etc.). Be sure to read and heed the notes at the bottom of the page. I recommend setting Capacitor Aging to 30%, and if you participate in distributive computing projects (e.g. BOINC or Folding@Home), I recommend setting TDP to 100%. Research your video card and pay particular attention to the power supply requirements for your card listed on your video card maker's website. If not listed, check a comparable card (same graphics engine and RAM) from a different maker. The key specifications, in order of importance are:

• Current (amperage or amps) on the +12V rail,
• Efficiency,
• Total wattage.

Then look for power supply brands listed under the "Good" column of PC Mechanic's PSU Reference List. Ensure the supplied amperage on the +12V rails of your chosen PSU meets the requirements of your video card. Don't try to save a few dollars by getting a cheap supply. Digital electronics, including CPUs, RAM, and today's advanced graphics cards, need clean, stable power. A good, well chosen supply will provide years of service and upgrade wiggle room. I strongly recommend you pick a supply with an efficiency rating equal to, or greater than 80%. Look for the 80 Plus - EnergyStar Compliant label. And don't forget to budget for a good UPS with AVR (automatic voltage regulation).

[Digerati]

Don't knock the technology - RDRAM is good stuff, if you can find it, and afford it - it was expensive then, outrageous now - but my i850 Gigabyte board with 2 x 512Mb 1066 RDRAM is still running strong - knock on forehead/wood.

So you have replaced the monitor and graphics card. You tested the RAM (did you let it run for several passes?) using MemTest. You might try Windows Memory Diagnostic - see the easy to follow instructions under Quick Start Information.

Disconnect any unneeded devices, especially external USB devices.

7. Power consumption - if it was a power problem there would be no way of overcoming the problem and the PC should crash when running a windowed game when the GFX card takes more power, but even plugging a second HDD doesn't make any difference.

Bad assumption. That would be the scenario for a seriously under-rated power supply, and hard drives consume very little (typically <20W when spinning up and <10W when idle) compared to other devices, especially graphics cards - which can easily out-pace CPUs. A stressed or failing supply, on the other hand, teeters on the threshold of stable/unstable operation, often causing very irregular (and difficult to troubleshoot) behavior in the devices it supplies power to, including the motherboard, while "appearing" to work fine.

Therefore, I ALWAYS start troubleshooting hardware by ENSURING I am feeding it quality power - most easily done by swapping in a known good PSU with sufficient horsepower. I see no mention of your PSU or voltages. What are the specs of your PSU? Is this the original? Has it been tested, or have you tried another? You MUST verify your power is (1) ample, and (2) stable (stays within ATX Form Factor standards - 5 or 10% depending on voltage) at ALL loads. It is a common mistake to ignore power requirements when adding or upgrading (or troubleshooting) hardware. This is especially critical when upgrading power hungry graphics cards.

Below is my canned text on testing PSUs, then another text on calculating power requirements and choosing a quality upgrade.

To properly and conclusively test a power supply unit (PSU), it must be tested under various realistic "loads" then analyzed for excessive [url="http://en.wikipedia.org/wiki/Ripple_(electrical)""]ripple[/url] and other anomalies. This is done by a qualified technician using an oscilloscope or power analyzer - sophisticated (and expensive) electronic test equipment requiring special training to operate, and a basic knowledge of electronics theory to understand the results. Therefore, conclusively testing a power supply is done in properly equipped electronic repair facilities.

Fortunately, there are other options. I keep a FrozenCPU Ultimate PSU Tester in my tool bag when I am "in the field" and don't have a good spare power supply to swap in. While not a certain test, they are better than nothing. The advantage of this model is that it has an LCD readout of the voltage. With an actual voltage readout, you have a better chance of detecting a "failing" PSU, or one barely within specified ATX Form Factor Standard tolerances. Lesser models use LEDs to indicate the voltage is just within some "range". These are less informative, considerably cheaper, but still useful for detecting PSUs that have already "failed". Newegg has several testers to choose from. All these testers contain a "dummy load" to fool the PSU into thinking it is connected to a motherboard, and therefore allows the PSU to power on, if able, without being attached to a motherboard - great for testing fans, but again, it is not a true load or suitable for conclusive testing.

As mentioned, swapping in a known good supply is a tried and trued method of troubleshooting used for centuries, even by pros. Remove the "suspect" part and replace with a "known good" part and see if the problem goes away.

I do not recommend using a multimeter to test power supplies. To do it properly, that is, under a realistic load, the voltages on all the pins must be measured while the PSU is attached to the motherboard and the computer powered on. This requires poking (with some considerable force) two hard and sharp, highly conductive meter probes into the main power connector, deep in the heart of the computer. One tiny slip can destroy the motherboard, and everything plugged into it. It is not worth the risk considering most multimeters, like plug-in testers, do not measure, or reveal any unwanted and potentially disruptive AC components to the DC voltages.

And remember, anything that plugs into the wall can kill. Do not open the power supply's case unless you are a qualified electronics technician. There are NO user serviceable parts inside a power supply.

***

Use the eXtreme PSU Calculator Lite to determine your power supply unit (PSU) requirements. Plug in all the hardware you think you might have in 2 or 3 years (extra drives, bigger or 2nd video card, more RAM, etc.). Be sure to read and heed the notes at the bottom of the page. I recommend setting Capacitor Aging to 30%, and if you participate in distributive computing projects (e.g. BOINC or Folding@Home), I recommend setting TDP to 100%. Research your video card and pay particular attention to the power supply requirements for your card listed on your video card maker's website. If not listed, check a comparable card (same graphics engine and RAM) from a different maker. The key specifications, in order of importance are:

• Current (amperage or amps) on the +12V rail,
• Efficiency,
• Total wattage.

Then look for power supply brands listed under the "Good" column of PC Mechanic's PSU Reference List. Ensure the supplied amperage on the +12V rails of your chosen PSU meets the requirements of your video card. Don't try to save a few dollars by getting a cheap supply. Digital electronics, including CPUs, RAM, and today's advanced graphics cards, need clean, stable power. A good, well chosen supply will provide years of service and upgrade wiggle room. I strongly recommend you pick a supply with an efficiency rating equal to, or greater than 80%. Look for the 80 Plus - EnergyStar Compliant label. And don't forget to budget for a good UPS with AVR (automatic voltage regulation).

[VarHyid]

Don't knock the technology - RDRAM is good stuff, if you can find it, and afford it - it was expensive then, outrageous now - but my i850 Gigabyte board with 2 x 512Mb 1066 RDRAM is still running strong - knock on forehead/wood.

That's what I've heard and why I've chosen RDRAM in the first place, then Rambus went to [bleep] and all I can find about RDRAM on the internet are complaints about it including its high power consumption (then read it's [bleep] so not sure it's true).

You tested the RAM (did you let it run for several passes?) using MemTest. You might try Windows Memory Diagnostic

One pass only. Will try the Windows Diag. later, thanks.

Disconnect any unneeded devices, especially external USB devices.

I have only one USB device which is my ADSL modem. Not really a power sucking monster No printers, no scanners, nothing.

Now for the power supply. You're absolutely right, that's something most people don't care about... and so did I By the time I was buying my (custom) PC, I didn't tell them what PSU should they put in assuming they know what they're doing and won't screw me on that and even if, then the other way around - by selling a better model especially when they saw what I'm capable of paying for it (and here we are at RDRAM again ). Now that I think of it, why didn't I get suspicious when the "chief technician" laughed at me once I told him RDRAM has to be mounted in pairs...

Anyway, that's what they've put inside (quote from the device):

HuntKey LW 8250 (v2.03)
300W (+5V & +3.3V MAX 150W)

Now let's check your power calculator... 4 RAMs... 2 DVD+/-RWs... TV card... SB+front bay... what do we have...
293W - Holy Sweet Mary Mother of Jesus Christ on a stick!!!

I've got 7 f***ing Watts free!?

OK, screw the one old DVD drive... 258W What!? It's an unused drive without any constantly spinning DVD inside.

BTW, it's only 219W with the default recommended settings (no Capacitor Aging and 85% TDP). I had no idea of "capacitor aging", but if it's true... [bleep], that sounds valid to me. This poor machine has being used ~10-15 hours a day for over 7 years

Unfortunately, I can't come up with a known good one to check if that's the problem and don't have the tools to check it. Maybe I'll find a new cheap (yeah, I know) PSU.

3 more questions:

1. Does an unused DVD+/-RW drive really suck over 20W just by being there? If so, the moment I bought a second drive (years ago) may indeed be the moment it started crashing. I'll get rid of it and post the results.

2. Does the size of RAM directly influence power consumption? Working with 512 MB (or 640 MB, for that matter) RAM was a blessing that day. FireFox up in a couple of seconds instead of 1/2 min, no HDD scratching, faster Flash animations.

3. Is there a possibility to test the PSU with some software? SiSoft Sandra tells me only the current voltages which are normal (within the allowed area)?

======== EDIT ========

Test 1, unplugging the drive - failed. Would be to good to be true.

Edited by VarHyid, 15 July 2009 - 11:37 AM.

[Digerati]

I appreciate the asterisks - but this is still a family friendly forum.

I had no idea of "capacitor aging", but if it's true...

What's current flow in a circuit but a whole bunch of electrons being jammed through a conductor at near light speed. No conductor is perfect so there is some friction and wasted energy, seen immediately as heat, and long term as wear and tear on the devices, reducing efficiency even further. Over time, they degrade enough to fail. Power supplies, inefficient to begin with, always lose power over time. This suggests to me your PSU is currently overstressed, and possibly not supplying your sensitive devices with "stable" voltages. This is often compounded by the fact the regulator circuits on the motherboard are years old too, and may no longer be able to compensate adequately.

All electronics, but especially motors, consume a very large amount of power when first starting up, but then settle down to a lessor rate of consumption. So when your DVD spins up, that would surely place a demand on the PSU.

Yes, more RAM, and more sticks consumes more power.

No, you cannot "test" a PSU with software - other than monitoring the voltages to ensure they remain stable under various loads. You need to swallow the pill and buy a new PSU or find some trusting soul to loan you a PSU for testing, or take it to a shop where it can be tested properly. This should be done NOW before that PSU fails catastrophically, and takes out everything connected to it.

Of course this may not be the PSU, and replacing it still may not fix your problem. And in that event, I would start looking suspiciously at the motherboard.

[VarHyid]

Done test #2 - Windows Memory Diagnostic. 1 standard test and 1 extended test - both passed without errors.

You need to swallow the pill and buy a new PSU or find some trusting soul to loan you a PSU for testing, or take it to a shop where it can be tested properly. This should be done NOW before that PSU fails catastrophically, and takes out everything connected to it.

I'll try to install WinXP on an empty partition to be 100% sure it's not "just" some OS/driver/software error, if this happens on a fresh new system - I'll try to find a new PSU. It's not like I don't want to spend a little money it's just... what money...

Of course this may not be the PSU, and replacing it still may not fix your problem. And in that event, I would start looking suspiciously at the motherboard.

That would be also "great".

[Digerati]

I'll try to install WinXP on an empty partition to be 100% sure it's not "just" some OS/driver/software error,

It's your time, but I note you have already tried changing the OS once.

1. OS - Started with the old Windows 98 SE, happens at Windows 2000 Professional as well.

[VarHyid]

Yes, but I didn't do it because of that and unfortunately, can't remember if it happened the first (and only) time I've changed the resolution right after installing Win2k and even if it didn't, maybe I just "got lucky".

Anyway, yep, my time and an excuse to check if it would work with 256 MB RAM

One more theoretical question - you've said power supply problems can cause irregular behavior in the device that gets some weird power... fluctuation. I'm anything else than an expert in electronics, but if the problem is caused by, let's say, a sudden change in the GFX-card's power consumption due to a resolution change, then why doesn't it EVER happen on startup when there are multiple resolution changes, incl. the big step from Windows loading screen to 1900x1200?

What are the odds that I always "get lucky" starting the system and (almost) always don't when once the system's up any application (incl. the system itself) fails to hold the new res. for more than a couple of seconds. Does the system suck that much more energy once Windows is fully loaded?

[Digerati]

then why doesn't it EVER happen on startup when there are multiple resolution changes

Perhaps because the CPU is not working hard. Or because having just rebooted, there was a drop in temperature in some device.

If I could answer your question, then I don't think it would be irregular problems - hence the difficulty in troubleshooting.